Factuality Detection on HotpotQA (test)
73.3AUROCDynHD
Evaluation Results
| Method | Links | |
|---|---|---|
| DynHDSource Dataset=TriviaQA, Method Category=Trajectory-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 73.3 | |
| TSVSource Dataset=TriviaQA, Method Category=Latent-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 65.5 | |
| TraceDetSource Dataset=CSQA, Method Category=Trajectory-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 65 | |
| DynHDSource Dataset=CSQA, Method Category=Trajectory-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 64.9 | |
| TSVSource Dataset=CSQA, Method Category=Latent-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 59.1 | |
| TraceDetSource Dataset=TriviaQA, Method Category=Trajectory-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 57.4 | |
| CCSSource Dataset=CSQA, Method Category=Latent-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 54 | |
| CCSSource Dataset=TriviaQA, Method Category=Latent-based, Evaluation Protocol=Zero-shot cross-task generalization2026.03 | 51.8 |